Template-Type: ReDIF-Paper 1.0
Series: Tinbergen Institute Discussion Papers
Creation-Date: 2000-02-10
Number: 00-006/4
Author-Name: J.S. Cramer
Author-Email: cramer@tinbergen.nl
Author-Workplace-Name: University of Amsterdam
Title: Asymptotic Properties of Predicted Probabilities in Discrete Regression
Abstract: The discrete outcome of a probability model is recordedas Y(i)=1 while otherwise Y(i)=0. y is the vector of observedoutcomes, p the corresponding probabilities, p^a consistent estimate of p, and residuals are defined ase = y - p^. Under quite general conditions, theasymptotic properties of p^ ensure that these residualshave zero mean and are uncorrelated with p^. Theseasymptotic results extend to the multinomial case. Theysupport certain measures of fit for discrete models.
File-Url: https://papers.tinbergen.nl/00006.pdf
File-Format: application/pdf
File-Size: 357955 bytes
Handle: RePEc:tin:wpaper:20000006